Scientific workflows and the computer applications that support them can be a complex topic. This book offers a good starting point. The introduction by the editors is well written, although it was obviously not adapted to the revised structure of this edition (for example, Part 1 is referenced in the book as “Background,” while the actual Part 1 of the book, “Application and User Perspective,” is referenced in the introduction as Part 2). Nevertheless, the introduction is a good starting point. There is a good comparison between scientific and business workflows, which will be useful for readers with a background in business processes.
Part 1, “Application and User Perspective” discusses a wide range of application areas (image mosaic generation, gravitational wave science, ecology, neuroscience, weather forecasting, and geology) and applied workflow tools (Pegasus, Directed Acyclic Graph Manager (DAGMan), message-passing interface (MPI) based implementations, Triana, Kepler, Condor). Some of these tools are described in Part 3.
Part 2, “Workflow Representation and Common Structure,” evaluates different ways to structure and describe scientific workflows. Included are both formal methods, such as graphs and Petri nets, and descriptive languages (including business process execution language (BPEL)), semantic notations, and other techniques to isolate physical representations of data from its logical typing.
Part 3, “Frameworks and Tools,” discusses workflow generation, refinement, and execution. Here the authors describe tools that were mostly already introduced as part of the implementation of the case studies in Part 1, or as proof of concept within Part 2
Although the editors have structured the contributions well, the book suffers from the typical syndrome of such works: the coherency of the different chapters is sometimes not sufficient. Nevertheless, the book is well suited as a starting point for readers who are searching for ways to define and execute workflows in their area of research. It is also a good starting point for scientists focused on further evolving the topic, both technically and methodically.